地圖概觀：
* x 軸是經度 ( Longitude ) ，y 軸是緯度 ( Latitude )

get_map() parameters

ggmap() parameters

# install.packages("ggmap")
# 如果安裝過程中遇到無法安裝png, jpeg 等問題 (macOS) 
# 請在terminal 執行 brew link libpng 和 brew link libjpeg，重試安裝
library(ggmap)
## Loading required package: ggplot2
library(mapproj)
## Loading required package: maps
map <- get_map(location = '台北市', zoom = 11, maptype = "roadmap")
## Map from URL : http://maps.googleapis.com/maps/api/staticmap?center=%E5%8F%B0%E5%8C%97%E5%B8%82&zoom=11&size=640x640&scale=2&maptype=roadmap&language=en-EN&sensor=false
## Information from URL : http://maps.googleapis.com/maps/api/geocode/json?address=%E5%8F%B0%E5%8C%97%E5%B8%82&sensor=false
ggmap(map)

ggplot2 的地圖

下面的範例利用前面練習 Data Visualization 的資料進行延伸，

library(maps)

# us.cities 包含 us 所有的country name, etc, population, lat, lon, capital
# 因為 txhousing 沒有經緯度資訊，所要利用 us.cities 找出 lat and lon
data(us.cities)

# Preprocessing
# collect all the city name in txhousing
mycountry <- unique(txhousing$city)

# beacause the us.cities$name is end in "TX", so we need to cut it off
tx.cities <- subset(us.cities, country.etc == "TX")
tx.cities$city <- unlist(strsplit(tx.cities$name, " TX"))

# for use rep to map the lat and long data conveniently, we choose only one year
m.txhousing <- subset(txhousing, year == 2000 & city %in% tx.cities$city)

# there are 12 months data in 2000, so set each = 12
temp <- tx.cities[tx.cities$city %in% m.txhousing$city, c("lat", "long")]
temp <- temp[rep(seq_len(nrow(temp)), each = 12), ]
m.txhousing.geo <- cbind(m.txhousing, temp)

# print data
head(m.txhousing.geo)
##        city year month sales   volume median listings inventory     date
## 1   Abilene 2000     1    72  5380000  71400      701       6.3 2000.000
## 1.1 Abilene 2000     2    98  6505000  58700      746       6.6 2000.083
## 1.2 Abilene 2000     3   130  9285000  58100      784       6.8 2000.167
## 1.3 Abilene 2000     4    98  9730000  68600      785       6.9 2000.250
## 1.4 Abilene 2000     5   141 10590000  67300      794       6.8 2000.333
## 1.5 Abilene 2000     6   156 13910000  66900      780       6.6 2000.417
##       lat   long
## 1   32.45 -99.74
## 1.1 32.45 -99.74
## 1.2 32.45 -99.74
## 1.3 32.45 -99.74
## 1.4 32.45 -99.74
## 1.5 32.45 -99.74
# draw plot
ggplot(m.txhousing.geo, aes(x = long, y = lat, size = sales, colour = cut(median, 5))) + # 設定ggplot圖資
  borders("county", "texas", colour = "grey70") + # 設定圖在德州，用county區分
  geom_point(alpha=.5) + # 顯示圖上的資料點
  facet_wrap(~month) +   # 區分出不同 month 的資料
  ggtitle("Housing market for populous cities in Texas (2000)") + # set plot title
  scale_colour_discrete(name  = "Median price") + # 設定不同顏色資料點的意義
  scale_size_continuous(name  = "Number of Sales") # 設定不同大小資料的意義
## Warning: Removed 37 rows containing missing values (geom_point).

ggmap + ggplot2

以下練習用 ggmap 結合 txhousing 的資料
* geocode: 可用來取得特定地區的經緯度資料，但免費版本僅限 2,500 次轉換

# get Texas's lat and lon
tx_center <- as.numeric(geocode("Texas"))
## Information from URL : http://maps.googleapis.com/maps/api/geocode/json?address=Texas&sensor=false
# 設定 ggmap
txMap <- ggmap(get_map(location = tx_center, zoom=6), extent="normal")
## Map from URL : http://maps.googleapis.com/maps/api/staticmap?center=31.968599,-99.901813&zoom=6&size=640x640&scale=2&maptype=terrain&language=en-EN&sensor=false
# 設定資料集，取得 Texas 的所有區塊
tx.cities.all <- subset(us.cities, country.etc == "TX")

# ggmapp 和 ggplot2 結合使用，可以在實體地圖上顯示資料
txMap + geom_point(aes(x=long, y=lat, size = pop), col = "orange", 
                   data = tx.cities.all, alpha=0.4) +
  ggtitle("Population of Texas cities")

References